SAFe POPM cheat sheet

Product Management is the content authority for the program (ART) backlog; the Product Owner is the content authority for the team backlog.

Product Management defines and prioritizes features with WSJF; the Product Owner decomposes features into stories with acceptance criteria and orders them.

Product Management faces outward (customers, market, ART stakeholders, roadmap); the Product Owner faces inward (team, iteration execution, day-to-day decisions).

One Product Manager typically serves an entire ART; one Product Owner typically serves one team, and may support two small teams.

The Product Owner accepts stories at the iteration level against acceptance criteria and the Definition of Done; feature-level validation happens at the system demo with Product Management.

Both roles attend PI Planning: Product Management presents the vision and top features, while the Product Owner facilitates team planning and objective drafting.

Product Management owns release content and program-level prioritization; the Product Owner owns iteration goals and backlog readiness for upcoming iterations.

The Product Owner does not define the vision unilaterally; vision, roadmap, and feature intent flow from Product Management and stakeholders into the team backlog.

Pre-PI work: Product Management refines, sizes, and prioritizes the program backlog using WSJF so the ordered feature set is ready before the event.

Readiness gate: each top feature has acceptance criteria, a size estimate, and known dependencies before teams can plan against it.

Teams compute PI capacity from the number of iterations, member availability, and known absences, then prepare draft PI objectives.

The Product Owner refines the team backlog so the highest-priority stories meet INVEST-style criteria and have acceptance criteria before planning.

Management prepares the business context briefing and the top features presentation that open the PI Planning event.

Logistics preparation covers attendees, location or virtual tooling, the agenda, the ART planning board, and the confidence-vote process.

Teams and Product Owners collect known risks, assumptions, and cross-team dependencies in advance so they can be placed on the planning board.

PI Planning inputs are the vision, the roadmap, the prioritized top features, team capacity, and draft objectives.

Business owners open the event by presenting business context and the vision, explaining the current state and the top features for the PI.

Draft plan review: each team presents its board, PI objectives, dependencies, and risks; leadership listens and asks questions rather than dictating team plans.

Management review and problem-solving follows the draft plan review, where scope, people, and dependencies are adjusted before the plan is finalized.

Confidence vote: every team member votes one to five fingers; a team average below three triggers discussion and re-planning until confidence is sufficient.

Risks are ROAMed: Resolved, Owned, Accepted, or Mitigated, with a named owner assigned for risks that are owned or mitigated.

Business owners assign business value from one to ten to each team's PI objectives, enabling later comparison of planned versus actual value.

The ART planning board makes cross-team dependencies visible; leaders help remove or reassign dependencies rather than leaving them for iteration execution.

Leadership behavior in PI Planning is decentralized decision-making: set context and constraints, then let teams own their plans and commitments.

Iteration planning: the team agrees an iteration goal, selects backlog items up to its capacity, and decomposes them into tasks to confirm feasibility.

Story points estimate relative size; velocity is the average number of points a team completes per iteration and is used for forecasting, not performance rating.

Team sync runs daily for about fifteen minutes and focuses on progress toward the iteration goal, coordination, and impediments.

The Product Owner accepts a story only when its acceptance criteria and the team's Definition of Done are satisfied.

Backlog refinement is continuous; roughly one to two iterations of stories should be ready before the next iteration planning.

The system demo occurs at the end of every iteration and shows integrated work from all teams in the ART, with stakeholder feedback captured.

An Innovation and Planning (IP) iteration follows each PI except the final one, providing time for innovation, backlog refinement, and planning without new story commitments.

Built-in quality practices - continuous integration, test-first, and automated testing - are applied inside the iteration rather than deferred to the end.

The iteration retrospective closes the iteration and produces improvement actions that the team tracks into the next iteration.

The ART sync is a weekly event combining the PO sync and the Scrum of Scrums so Product Management, Product Owners, and teams align on PI objective progress.

PO sync reviews progress toward PI objectives, backlog changes, and scope adjustments; it does not replace team-level iteration events.

A system demo runs each iteration; the PI System Demo at PI end demonstrates cumulative integrated value from all teams.

Inspect and Adapt follows the PI System Demo and includes quantitative and qualitative measurement plus a structured problem-solving workshop.

Program predictability is measured as the business value of committed PI objectives actually achieved versus the value committed.

Uncommitted PI objectives are planned but excluded from the commitment, preserving capacity for emergent work.

Release on demand decouples releases from the PI cadence; the continuous delivery pipeline supports deploying and releasing when business needs dictate.

Continuous delivery pipeline, DevOps, and built-in quality practices are the technical enablers that let the ART deliver value during PI execution.

Generative AI can draft story descriptions, acceptance criteria, and test ideas from feature context; the Product Owner reviews and corrects every draft before it enters the backlog.

AI summarizes and clusters large volumes of customer feedback, reviews, and support tickets so Product Management can surface themes faster than manual triage.

AI can inform WSJF inputs through demand forecasting, sentiment analysis, and effort estimation, but people make and own the prioritization decisions.

Guardrails rule: no confidential, personal, or proprietary data goes into an unapproved AI tool; use only tools sanctioned by organizational policy.

AI output is treated as a draft: verify factual accuracy, check for bias or fabricated detail, and confirm alignment with product intent before use.

Accountability for AI-assisted artifacts and decisions stays with the Product Owner or Product Manager, not with the tool.

AI can accelerate backlog refinement by proposing story splits and dependency candidates, but the team validates estimates and technical feasibility.

Document where AI assisted significant decisions so the reasoning remains auditable under organizational policy.